Author: J. Allen Byrne
Switched Mode Rectifier/chargers (SMR’s) are the dominant method of stationary battery charging for telecommunications, information technology (IT) and other standby power application. The vast majority of these are forced air cooled. In North America, the exception is with the utility companies who still favor older linear technologies, mainly Silicone Controlled Rectifiers (SCR’s).
So why the reluctance of utility companies to adopt the lighter, more efficiuser-friendlyfriendly SMR technology? This presentation will explore the issues and argue the case for a second look at SMR’s, in particular, convection cooled SMR’s.
The author uses his experience in working with all the major user groups and also with the various charger technologies, having first becoming familiar with SMR chargers with their popular acceptance in the late 1980’s. He understands the reasons that utility users are reluctant to change and thinks that many industrial and utility users are “missing the boat” by not adopting modular power systems using modular SMR technology and the benefits they bring. The objective of this presentation is to be an educational thought changer.
Topics discussed include; history, reliability, efficiencies, maintenance, pros and cons, and costs.
The author, before his recent retirement, spent a good part of the previous five years working closely with utility companies and their battery back-up systems. This work included sizing, configuring, surveying, maintaining, and installing both new and replacement systems. Having worked for many years with communications and information technology users, he kept wondering why the utilities used a different approach to providing battery back-up. This paper is a brief discussion of that wonderment.
A Brief History
First of all, a little history of battery rectifiers/chargers. This is the preferred name as these devices have to perform both functions, and that is, rectify alternating current (ac) to direct current (dc) and at the same time be capable of charging a battery. When stationary batteries are in use, a charge has to be maintained on them whether it is to compensate for self-discharge or to recharge the battery after discharge. The dc output of the rectifier charger must be well filtered and regulated to be able to do this without having a detrimental effect on the battery.
The Mag Amp
Early in the last century a device was invented to magnify electrical signals and was called a magnetic amplifier which came to be used in place of vacuum tube amplifiers as they were more robust. This “Mag Amp” as it became to be known, was a solid state device and with high reliability and it became the workhorse for battery charging. Indeed these high-reliability mag-amp based chargers are still being used today.
The Controlled Ferro
The 1930’s saw another transformer design in which the output voltage could be maintained within the required limits irrespective of the input voltage or output load. In this case it was the physical design of the transformer that did the control, not any external feedback. It was possible however to vary the output voltage with an external circuit and this is the basis of what we know as a Controlled Ferroresonant Rectifier/charger (CFR).
During the 1950’s, Bell Laboratories developed the Silicone Controlled Rectifier (SCR) and this was used along with conventional transformers to come up with a new rectifier/charger technology which became known as the SCR charger. This technology allowed the output to the battery to be electronically controlled rather than through the design of the transformer. This also allowed for simpler and cheaper transformers to be employed. This basic technology with many refinements along the way, such as microprocessor control and communications abilities, became and still is, the workhorse of the utility and industrial market segment today. However the magnetic amplifier and ferroresonant transformer based designs are still being manufactured and used in specific applications.
The Times Were a Changing
During the late 1970’s and early 1980’s, things were happening in the deployment of telecommunications (now usually referred to as communications) systems and the computing industries. There were also new developments in battery technologies. These things combined spurned a new approach to battery charging.
On the communications side, In the USA, the Bell System, which was owned by AT&T, had a virtual monopoly of providing telephone service. Because of this total control and public resentment, an antitrust lawsuit was filed, United States v. AT&T, in an attempt to break up this monopoly. A compromise was reached by AT&T and as part of the settlement of this legal action, on January 1, 1984, the Bell System was broken up into smaller Regional Bell Operating Companies (RBOC’s) or “Baby Bells.” In the rest of the world, communications companies were mainly government controlled, for example, British Telecom and France Telecom. Because of the settlement there were several competitors waiting in the wings such as MCI and Sprint and other manufacturers of communications switching equipment such as ROLM. One of the main outcomes was that the RBOC’s no longer had to buy equipment including battery backup systems from AT&T. Also, the new start-up communications providers were able build up their networks with more modern and innovative equipment. As a result, the industry was moving away from large, centralized, switching centers and deploying more and more network equipment outside of the large central offices.
In tandem with the developments in the communications industry, the IT industry was also experiencing change. Large computer mainframes and their associated support equipment was shrinking and powering techniques and demands were changing. Distributed computing was also being used and because of the increase in processing capabilities, smaller computer systems were being installed in locations other than dedicated, large computer rooms.
Some other things were also happening and they also impacted powering methods and battery backup techniques. There was the rapid development in deploying cellular (mobile) radio. During the late 1980’s and early 90’s this wireless technology was being deployed at a rapid rate and battery backup system
providers where anxious to keep pace. The same applied to the cable TV deployment with cable head ends and networks being built.
The Perfect Storm
All of these things happening basically at the same time, created in the author’s opinion, the equivalent of the “perfect storm” in the battery backup field and the author lived and worked through it.
So how does these happenings affect battery charging and batteries? It was basically customer driven. Because of the remote locations, residential locations, customer premise locations (communications equipment being located at the users’ premises), etc., users were asking for smaller, lighter, cheaper, more user friendly and easily maintainable equipment. Well, it just so happened that there were two other break through innovations that were happening at the same time.
There was a move on both sides of the Atlantic to come up with a more user-friendly lead-acid battery, one that did not have liquid electrolyte or was subject to out-gassing. In Europe, the first gelled electrolyte battery was developed by Sonnenschein in Germany. The first Absorbed Glass Mat (AGM) battery was developed by Gates Rubber Company in the USA. A number of other manufacturers jumped on the wagon including Chloride and Tungstone in the UK and Gould National Batteries (GNB) in the USA. They developed and refined the Valve-Regulated Lead-Acid (VRLA) battery we know today. In the UK this was stimulated by British Telecom specifying this type of battery to support their new communications technologies. In North America, GNB took a lead in developing larger format units and marketing these VRLA Batteries. However, there were many teething problems and customer satisfaction was poor. Indeed, much of this led to the formation of Battcon™ in 1995.
In a similar manner, much of the development of rectifier charging technologies was user driven. Just like with VRLA batteries, customers wanted something smaller, lighter, cheaper, more user friendly and easily maintainable They also wanted equipment that was less noisy, both electrically and physically than the existing CFR and SCR equipment. This was the introduction to Switched Mode Rectifier (SMR) battery rectifier/chargers. A development in the late 1950’s of a method in which the ac input to the rectifier was converted to a much higher frequency using switching transistors which allows smaller transformers to be used. The SMR charger was the answer to the user demands. Certainly, they were smaller, lighter, more user friendly, quieter and in a modular configuration, easier to maintain. However, they had some early drawbacks in that they were still quite bulky when convection cooled and the earlier versions were not as efficient as they are today, and managing heat dissipation required sizeable heat sinks. The answer was to make them forced air cooled. This certainly made the SMR’s smaller by eliminating the large heat sinks but by introducing fans, that caused some other problems. The fan became the highest single point of failure and also through the ingestion of dust and dirt and clogged filters caused other problems, and in addition, the fan load affected efficiency. A development called “fan-on-demand” where the fan was only activated if the unit temperature rise requires the fan to run. Fan reliability has also improved over the years and efficiency has also improved, with some having a life-span approaching 20 years. Fan cooled SMR’s now serve the majority of battery charging requirements. However, there is one exception which leads, albeit in a somewhat lengthy way, to the main point of this paper. That exception is the use of SMR’s in industrial and utility application.
The Reluctance to Change
So why was this market segment reluctant to move with the times? There were several reasons. For a start, those industries were traditionally reluctant to change and preferred to stick with what worked. There was little or no change in the equipment that was being powered nor was there load equipment downsizing or change of location outside of the controlled environs that they operated in. There was no exposure to non-qualified personnel. Space was not a problem so there was no need to make things smaller, lighter, etc. There was usually plenty of manpower and technicians to maintain the equipment. There was no move to use the new technology VRLA batteries as they rightly preferred to stay with the proven reliability of Vented Lead-Acid (VLA) batteries.
SCR chargers have over the years, proved to be a very reliable means of charging batteries; however, SCR chargers had and have some drawbacks. They as not as efficient as SMR’s and the cost of powering and efficiency is being scrutinized by regulatory agencies. Regulated utilities may be exempt, but for how long? One example is the State of California, which in 2011 mandated in Section 25402, subdivision (c), of the Public Resources Code1 that the “California Energy Commission reduce the wasteful, uneconomic, inefficient, or unnecessary consumption of energy by prescribing standards for minimum levels of operating efficiency of appliances whose use, as determined by the Commission, requires a significant amount of energy on a statewide basis.” At the point in time, the mandate only applies to devices under 2kW but watch for this to change. The author believes that the states of Oregon and Washington also have similar mandates. Almost all modern SMR’s would meet this mandate but that is not to say SCR’s would not but the fact is that SMR’s are more efficient. See Figure 1 and Table 1.
Figure 1 below is a graphic representation of the difference in efficiencies between a typical SCR charger and a SMR charger.
This graph represents two 130 Volt 30 Amp rectifier/chargers, which equals 3,900 Watts output. At an 85% efficiency at full load, the SCR charger would require 4,588 Watts ac input. The SMR charger at full load at 91% efficiency would only require 4,285 Watts ac input. This translates to a difference of 303 Watts which is 2,654 kW per year. At a generous cost of 10 US cents per KW/hr., this represents a cost savings of $265 per year or more realistically, $2,650 over a 10 year life span. This does not include the cost savings related to maintenance and repair.
Table 1 shows the results of an actual laboratory test of a SCR charger and a SMR charger of similar output capacities. The SMR charger, like the SCR unit, was convection cooled. This test clearly verifies the assumptions detailed in the example under Figure 1 above. At full load, the efficiency difference is 400 Watts which is 3,504 kWhr/year. Bur realistically, the rectifier/chargers are not going to be running at full load, so taking the example of the 4 amp load in the table above, the difference in efficiency is only 116 Watts which is 1,016 kWhr/year. which is still a significant power saving. As stated by the Electric Power Research Institute (EPRI) 2 “Switch mode chargers of any topology are generally more efficient than the other types of charger.”
Another problem with SCR chargers is that they sometimes can generate unsmoothed dc voltages causing a higher ripple voltage and current to be imposed on the battery. It is not fully understood exactly what all the effects of this ripple is on the battery, especially VRLA batteries, but battery heating is thought to be a concern. SCR chargers are typically connected to VLA batteries which are more tolerant to ripple but what if the industrial industries are persuaded to use the more environmentally friendly VRLA batteries. Don’t think that this is not possible, as there has already been some moves in this direction. Indeed, some regions and countries are attempting to do away with lead based batteries entirely. That is certainly going to have an impact on stationary battery charging but that is the subject matter for another paper.
Since rectifier/charger size or the space that is occupied doesn’t appear to be a problem in the industrial and utility industries and the perception that fan cooled SMR’s have reliability concerns, why not revert to the original type of convection cooled SMR units that were first utilized in the 1970’s and 1980’s? They were bulky but not as large as SCR chargers and didn’t have the communications and microprocessor control capabilities of today’s SCR chargers. So, if these capabilities were added to a
modern, high reliability, convection cooled SMR chargers, and they vaguely looked like a SCR charges, and they would operate in industrial environments, and had the reliability equal to or superior to that of SCR chargers, could they be accepted by the industrial/utility segment? The author believes that they would.
Typical industrial and utility installations that the author has seen, consist of a single SCR charger supporting a single VLA battery string. A few have dual SCR chargers, either individually connected to the dc bus or load sharing. Only a few have VRLA batteries, some of which are two or more parallel strings.
Where there is only a single charger the risk is obvious. If that single source of power fails for any reason, then the load is going to be running on battery. A well-engineered system will probably have only about 8 hours of battery reserve so unless there is a spare charger on site, the probability is that it will take more than 8 hours to fix or replace that charger. If it has to be replaced, because of the weight and for safety reasons, it would typically be a two man job. The risk to the load being compromised is obvious. The author, coming primarily from a communications and IT background, could not understand why there was not more redundancy built in to the overall system. There is usually a similar situation with the battery if only a single string is installed. If it has to be taken off line for any reason and there is a commercial power aberration, the load is at risk. Shouldn’t a well-engineered, fault tolerant power system have the following features? Dual ac inputs to the rectifier/charger(s), redundant paralleled rectifier/chargers, two or more battery strings connected in parallel, and the ability to operate if there is a failure in the control electronics. Other redundancy features that eliminate a single point of failure are outside the scope of this paper.
The ace in the hole when it comes to SMR chargers is that they are easily parallelable and can be packaged into a modular redundant configuration. Depending upon the size of the load, including battery recharge, a SMR system can be configured so that the total charging capacity can be supported by one or more modular rectifier/chargers and an additional unit added to make the system redundant. Here is a typical example of a scenario. Say the load including battery charging requires 20 amps. If a SCR charger is to be used then the rectifier/charger must be at least capable of supplying 20 amps. If redundancy is required, then another 20 amp unit must be installed. If a forced load sharing feature is not available then it is possible that one rectifier/charger will be doing all the work and will probably fail before the other. In addition, depending at which point in the efficiency curve they are individually running, overall efficiency may suffer. If for some reason the load increases such as with the addition of more equipment, then a third 20 amp rectifier charger must be added in order to maintain redundancy.
If a modular, convection cooled approach using a SMR charger is taken for the same 20 amp load, it could be configured as follows. Say the modular rectifier/charger is capable of 4 amps output, then 20 amps ÷ 4 amps = 5, which means that five 4 amp units would be required to supply the load. In order to supply redundancy, only one more 4 amp unit would be required and not another 20 amp unit as for the SCR charger example. If the cabinet housing the modular SMR chargers is so designed, it could accommodate two or more ac inputs and dc outputs. With suitable onboard electronics, forced load sharing and other features such as remote battery sensing and temperature compensation would be afforded.
Another big advantage in using the modular SMR charger approach is the serviceability. Since all the modular power units are hot swappable, if one control module fails it can be replaced with a new unit within a few seconds by a relatively unskilled person. This also means that the rectifier/charger system remains in full operation, supporting the load during the repair. This unit swap out can also be done without any exposure to live parts, a possible arc flash incident, or electrical shock.
Although, not so important in industrial applications, the convection cooled SMR charger is acoustically quite, whereas the SCR charger is inherently audibly noisy because of the transformers and chokes and this noise can increase as the SCR charger ages.
Lastly, the author took a look at pricing or cost per watt and it is very much dependent upon the electrical size of the system, in most cases it was pretty much parity. The same is true for energy density.
While SCR chargers are very well established with a proven reliability, there are some drawback, mainly with regards to the cost of redundancy, maintainability, scalability and servicing. It is the author’s opinion that when all the pros and cons are considered, a battery backed power system supported by SMR chargers is superior in most ways to a power system based upon SCR chargers. The author has no vested interest in either approach and much of the data he has accessed is from companies that manufacture and market both types of rectifier/chargers.
Download associated powerpoint presentation here: switch-mode-chargers-presentation
1.California Energy Commission. 2011 Appliance Efficiency Rulemaking. October 2011.
2.Geist, T. Et al. Designing Battery Charger Systems for Improved Energy Efficiency. Electric PowerResearch Institution Solutions, Inc. Prepared for the California Energy Commission.